Achievements in European Research on Grid Systems (eBook)

Contents 5
Foreword 7
Contributing Authors 10
DIVIDE ET IMPERA: PARTITIONING UNSTRUCTURED PEER- TO- PEER SYSTEMS TO IMPROVE RESOURCE LOCATION 18
1. Introduction 19
2. RelatedWork 20
3. The Partitions Design 22
4. Experimental Results 24
5. Conclusions 28
Acknowledgments 28
References 28
VALIDATING DESKTOP GRID RESULTS BY COMPARING INTERMEDIATE CHECKPOINTS 30
1. Introduction 31
2. Assumptions and Definitions 33
3. Comparison of Equivalent Checkpoint Digests 34
3.1 Reducing the Time to Detect an Error 34
3.2 Theoretical Analysis 34
4. Checkpoint-based Task Replication 37
5. Experimental Results 38
6. RelatedWork 39
7. Conclusion 40
References 41
INTEGRATION OF THE ENANOS EXECUTION FRAMEWORK WITH GRMS* 42
1. Introduction 43
2. eNANOS Execution Environment 43
3. Grid Resource Management System (GRMS) 45
4. Integration Issues 46
4.1 eNANOS as a new execution framework 48
4.2 eNANOS as a new information provider 52
5. Conclusions and Future Work 55
References 56
USER-TRANSPARENT SCHEDULING FOR SOFTWARE COMPONENTS ON THE GRID 57
1. Introduction 58
2. Context and Background 59
2.1 Environment Description 59
2.2 Higher-Order Components 60
2.3 User-Transparent Scheduling for HOCs 61
3. Implementation of the Scheduler 62
3.1 Integration of Existing Systems 62
3.2 Performance Evaluation of Scheduling 62
4. KOALA-based User-Transparent Scheduling Evaluation 64
4.1 Integration Feasibility 64
4.2 Performance Results 65
5. RelatedWork 66
6. Conclusions 67
Acknowledgments 68
References 68
PROBLEM SOLVING ENVIRONMENT FOR DISTRIBUTED INTERACTIVE APPLICATIONS 70
1. Introduction 71
2. PSEs for distributed interactive applications 71
2.1 Computational Steering Environment 72
2.2 CUMULVS 72
2.3 Cactus Problem Solving Environment 72
2.4 Discover 73
2.5 TENT 74
2.6 Interactive Simulation Systems Conductor 74
2.7 Comparision of existing PSEÌs 74
3. PSE for HLA-based simulations 76
3.1 Need for a Grid for HLA-based applications 76
3.2 Grid HLA Management System 77
4. Application of G-HLAM to vascular reconctruction 78
5. Summary and future work 79
Acknowledgments 80
References 80
FAULT-TOLERANT DATA SHARING FOR HIGH- LEVEL GRID PROGRAMMING: A HIERARCHICAL STORAGE ARCHITECTURE 82
1. Introduction 83
2. Analysis: using JuxMem to enable grid-level, fault-tolerant storage in ASSIST 84
2.1 Data sharing in ASSIST 84
2.2 Existing building blocks: ad-HOC and JuxMem 85
3. Proposal: an integrated 2-tier architecture 88
4. ASSIST and fault-tolerance: a sample scenario 90
5. Design and Implementation 91
5.1 Preliminary Experiments 92
6. Conclusion 94
References 95
PAL: EXPLOITING JAVA ANNOTATIONS FOR PARALLELISM 97
1. Introduction 98
2. Parallel Abstraction Layer (PAL) 99
3. A PAL prototype 102
4. Experimental results 104
5. Related work 106
6. Conclusion and future work 107
References 109
A NEW APPROACH ON NETWORK RESOURCES MANAGEMENT IN GRIDS* 111
1. Introduction 112
2. Grid Resource Broker 113
3. Network Resource Manager and Information Services 114
3.1 Network Resource Management System 115
3.2 Network Configuration Management System 116
3.3 GNRB and Multi-Point Real-Time Network Monitoring System 116
4. Network-aware Grid Resource Broker 116
5. Experimental Evaluation 119
6. Conclusion 121
References 121
COMPONENTISING A SCIENTIFIC APPLICATION FOR THE GRID* 123
1. Introduction 124
2. Background on Jem3D 124
2.1 The ProActive library 124
2.2 Jem3D overview 125
3. Approach 126
3.1 Componentisation process 127
3.2 Fractal/ProActive component model 129
4. Componentising Jem3D 130
5. Performance results 132
6. Related work 133
7. Conclusion 134
References 135
A PEER-TO-PEER FRAMEWORK FOR RESOURCE DISCOVERY IN LARGE- SCALE GRIDS 137
1. Introduction 138
2. Related work 139
3. Resources and query types 141
4. System architecture 144
4.1 Local components 144
4.2 Static attribute discovery 145
4.3 Dynamic attribute discovery 146
5. Conclusions 149
Acknowledgements 150
References 150
GRID SUPERSCALAR AND GRICOL: INTEGRATING DIFFERENT PROGRAMMING APPROACHES 152
1. Introduction 153
2. GriCoL 154
3. GRID superscalar 157
4. Studying the integration possibilities 160
4.1 Control-flow level integration 160
4.2 Data-flow level integration 161
4.3 General integrations 162
5. Conclusions and future work 163
Acknowledgments 163
References 163
DERIVING POLICIES FROM GRID SECURITY REQUIREMENTS MODEL* 164
1. Introduction 165
2. Security Requirements Model 165
2.1 Problem Statement 165
2.2 Goal Model 166
2.3 Responsibility Model 166
2.4 Object Model 167
2.5 Operation Model 168
2.6 Dealing with Obstacles 168
3. Derivation of Policies from Security Requirements Model 169
3.1 Refinement of Requirements Model 171
3.2 Policy Templates 172
3.3 Population of Policy Templates 173
3.4 Refinement of Policies 173
3.5 Implementation of Policies 174
4. RelatedWork and Discussions 175
5. Conclusions and Future Directions 175
References 176
DOMAIN-SPECIFIC METADATA FOR MODEL VALIDATION AND PERFORMANCE OPTIMISATION 177
1. Introduction 178
2. GENIE Application 180
2.1 Overview 180
2.2 The Past, the Present and the Future 181
3. Motivation 181
4. Metadata for Legacy Components 182
5. Staging the Metadata 183
5.1 Model Validation 183
5.2 Performance Optimisation 184
6. RelatedWork 185
7. Conclusions 186
References 188
A SERVICE FOR RELIABLE EXECUTION OF GRID APPLICATIONS 190
1. Introduction 191
2. Failures in Grid Environments 191
2.1 Related Work 192
3. A Service for Reliable Application Execution 193
3.1 Design objectives 193
3.2 Service design 195
3.3 Application wrapping 196
3.4 Failures detected by the service 197
3.5 Recovery techniques 198
3.6 Evaluation 199
4. Case Study: Reliable Execution for GRID superscalar 200
5. Conclusions 202
Acknowledgments 202
References 202
PERFORMANCE MONITORING OF GRID SUPERSCALAR WITH OCM- G/ G- PM: INTEGRATION ISSUES 204
1. Introduction 205
2. Monitoring requirements for and from the GS environment 206
3. Current status of research 207
4. The G-PM performance tool and its enhancements 209
5. Other integration issues 210
6. Case Study 211
7. Visualization of GS applications with GSM 213
8. Summary 214
References 215
IMPROVING WORKFLOWEXECUTION THROUGH SLA- BASED ADVANCE RESERVATION 217
1. Introduction 218
1.1 Related work 218
1.2 Remainder of the Paper 219
2. Service Level Agreements - State of the art 220
2.1 Technology Overview 220
2.2 Defining Service Level Objectives and Penalties 221
2.3 WS-Agreement 222
3. SLA based service provisioning for workflows 223
3.1 The workflow resource problem 223
3.2 Negotiating resource usage 224
3.3 MSS Implementation 225
4. Results 226
5. Future Perspectives 228
Acknowledgments 229
References 229
DEPENDABILITY EVALUATION OF THE OGSA- DAI MIDDLEWARE 232
1. Introduction 233
2. RelatedWorks 233
3. OGSA-DAI Overview 234
4. QUAKE: A Benchmarking Tool for Grid Services 236
5. Experimental Results 237
5.1 Performance Overhead 237
5.2 Benchmarking Tomcat+Axis 238
5.3 Benchmarking OGSA-DAI 239
6. Conclusions 243
Acknowledgements 244
References 244
Author Index 246